2 * Note: this file was generated by the Gromacs c kernel generator.
4 * This source code is part of
8 * Copyright (c) 2001-2012, The GROMACS Development Team
10 * Gromacs is a library for molecular simulation and trajectory analysis,
11 * written by Erik Lindahl, David van der Spoel, Berk Hess, and others - for
12 * a full list of developers and information, check out http://www.gromacs.org
14 * This program is free software; you can redistribute it and/or modify it under
15 * the terms of the GNU Lesser General Public License as published by the Free
16 * Software Foundation; either version 2 of the License, or (at your option) any
19 * To help fund GROMACS development, we humbly ask that you cite
20 * the papers people have written on it - you can find them on the website.
28 #include "../nb_kernel.h"
29 #include "types/simple.h"
34 * Gromacs nonbonded kernel: nb_kernel_ElecRF_VdwBham_GeomW4P1_VF_c
35 * Electrostatics interaction: ReactionField
36 * VdW interaction: Buckingham
37 * Geometry: Water4-Particle
38 * Calculate force/pot: PotentialAndForce
41 nb_kernel_ElecRF_VdwBham_GeomW4P1_VF_c
42 (t_nblist * gmx_restrict nlist,
43 rvec * gmx_restrict xx,
44 rvec * gmx_restrict ff,
45 t_forcerec * gmx_restrict fr,
46 t_mdatoms * gmx_restrict mdatoms,
47 nb_kernel_data_t * gmx_restrict kernel_data,
48 t_nrnb * gmx_restrict nrnb)
50 int i_shift_offset,i_coord_offset,j_coord_offset;
51 int j_index_start,j_index_end;
52 int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
53 real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
54 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
55 real *shiftvec,*fshift,*x,*f;
57 real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
59 real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
61 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
63 real ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
65 real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
66 real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
67 real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
68 real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
69 real dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
70 real velec,felec,velecsum,facel,crf,krf,krf2;
73 real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
82 jindex = nlist->jindex;
84 shiftidx = nlist->shift;
86 shiftvec = fr->shift_vec[0];
87 fshift = fr->fshift[0];
89 charge = mdatoms->chargeA;
95 vdwtype = mdatoms->typeA;
97 /* Setup water-specific parameters */
99 iq1 = facel*charge[inr+1];
100 iq2 = facel*charge[inr+2];
101 iq3 = facel*charge[inr+3];
102 vdwioffset0 = 3*nvdwtype*vdwtype[inr+0];
107 /* Start outer loop over neighborlists */
108 for(iidx=0; iidx<nri; iidx++)
110 /* Load shift vector for this list */
111 i_shift_offset = DIM*shiftidx[iidx];
112 shX = shiftvec[i_shift_offset+XX];
113 shY = shiftvec[i_shift_offset+YY];
114 shZ = shiftvec[i_shift_offset+ZZ];
116 /* Load limits for loop over neighbors */
117 j_index_start = jindex[iidx];
118 j_index_end = jindex[iidx+1];
120 /* Get outer coordinate index */
122 i_coord_offset = DIM*inr;
124 /* Load i particle coords and add shift vector */
125 ix0 = shX + x[i_coord_offset+DIM*0+XX];
126 iy0 = shY + x[i_coord_offset+DIM*0+YY];
127 iz0 = shZ + x[i_coord_offset+DIM*0+ZZ];
128 ix1 = shX + x[i_coord_offset+DIM*1+XX];
129 iy1 = shY + x[i_coord_offset+DIM*1+YY];
130 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
131 ix2 = shX + x[i_coord_offset+DIM*2+XX];
132 iy2 = shY + x[i_coord_offset+DIM*2+YY];
133 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
134 ix3 = shX + x[i_coord_offset+DIM*3+XX];
135 iy3 = shY + x[i_coord_offset+DIM*3+YY];
136 iz3 = shZ + x[i_coord_offset+DIM*3+ZZ];
151 /* Reset potential sums */
155 /* Start inner kernel loop */
156 for(jidx=j_index_start; jidx<j_index_end; jidx++)
158 /* Get j neighbor index, and coordinate index */
160 j_coord_offset = DIM*jnr;
162 /* load j atom coordinates */
163 jx0 = x[j_coord_offset+DIM*0+XX];
164 jy0 = x[j_coord_offset+DIM*0+YY];
165 jz0 = x[j_coord_offset+DIM*0+ZZ];
167 /* Calculate displacement vector */
181 /* Calculate squared distance and things based on it */
182 rsq00 = dx00*dx00+dy00*dy00+dz00*dz00;
183 rsq10 = dx10*dx10+dy10*dy10+dz10*dz10;
184 rsq20 = dx20*dx20+dy20*dy20+dz20*dz20;
185 rsq30 = dx30*dx30+dy30*dy30+dz30*dz30;
187 rinv00 = gmx_invsqrt(rsq00);
188 rinv10 = gmx_invsqrt(rsq10);
189 rinv20 = gmx_invsqrt(rsq20);
190 rinv30 = gmx_invsqrt(rsq30);
192 rinvsq00 = rinv00*rinv00;
193 rinvsq10 = rinv10*rinv10;
194 rinvsq20 = rinv20*rinv20;
195 rinvsq30 = rinv30*rinv30;
197 /* Load parameters for j particles */
199 vdwjidx0 = 3*vdwtype[jnr+0];
201 /**************************
202 * CALCULATE INTERACTIONS *
203 **************************/
207 c6_00 = vdwparam[vdwioffset0+vdwjidx0];
208 cexp1_00 = vdwparam[vdwioffset0+vdwjidx0+1];
209 cexp2_00 = vdwparam[vdwioffset0+vdwjidx0+2];
211 /* BUCKINGHAM DISPERSION/REPULSION */
212 rinvsix = rinvsq00*rinvsq00*rinvsq00;
213 vvdw6 = c6_00*rinvsix;
215 vvdwexp = cexp1_00*exp(-br);
216 vvdw = vvdwexp - vvdw6*(1.0/6.0);
217 fvdw = (br*vvdwexp-vvdw6)*rinvsq00;
219 /* Update potential sums from outer loop */
224 /* Calculate temporary vectorial force */
229 /* Update vectorial force */
233 f[j_coord_offset+DIM*0+XX] -= tx;
234 f[j_coord_offset+DIM*0+YY] -= ty;
235 f[j_coord_offset+DIM*0+ZZ] -= tz;
237 /**************************
238 * CALCULATE INTERACTIONS *
239 **************************/
243 /* REACTION-FIELD ELECTROSTATICS */
244 velec = qq10*(rinv10+krf*rsq10-crf);
245 felec = qq10*(rinv10*rinvsq10-krf2);
247 /* Update potential sums from outer loop */
252 /* Calculate temporary vectorial force */
257 /* Update vectorial force */
261 f[j_coord_offset+DIM*0+XX] -= tx;
262 f[j_coord_offset+DIM*0+YY] -= ty;
263 f[j_coord_offset+DIM*0+ZZ] -= tz;
265 /**************************
266 * CALCULATE INTERACTIONS *
267 **************************/
271 /* REACTION-FIELD ELECTROSTATICS */
272 velec = qq20*(rinv20+krf*rsq20-crf);
273 felec = qq20*(rinv20*rinvsq20-krf2);
275 /* Update potential sums from outer loop */
280 /* Calculate temporary vectorial force */
285 /* Update vectorial force */
289 f[j_coord_offset+DIM*0+XX] -= tx;
290 f[j_coord_offset+DIM*0+YY] -= ty;
291 f[j_coord_offset+DIM*0+ZZ] -= tz;
293 /**************************
294 * CALCULATE INTERACTIONS *
295 **************************/
299 /* REACTION-FIELD ELECTROSTATICS */
300 velec = qq30*(rinv30+krf*rsq30-crf);
301 felec = qq30*(rinv30*rinvsq30-krf2);
303 /* Update potential sums from outer loop */
308 /* Calculate temporary vectorial force */
313 /* Update vectorial force */
317 f[j_coord_offset+DIM*0+XX] -= tx;
318 f[j_coord_offset+DIM*0+YY] -= ty;
319 f[j_coord_offset+DIM*0+ZZ] -= tz;
321 /* Inner loop uses 157 flops */
323 /* End of innermost loop */
326 f[i_coord_offset+DIM*0+XX] += fix0;
327 f[i_coord_offset+DIM*0+YY] += fiy0;
328 f[i_coord_offset+DIM*0+ZZ] += fiz0;
332 f[i_coord_offset+DIM*1+XX] += fix1;
333 f[i_coord_offset+DIM*1+YY] += fiy1;
334 f[i_coord_offset+DIM*1+ZZ] += fiz1;
338 f[i_coord_offset+DIM*2+XX] += fix2;
339 f[i_coord_offset+DIM*2+YY] += fiy2;
340 f[i_coord_offset+DIM*2+ZZ] += fiz2;
344 f[i_coord_offset+DIM*3+XX] += fix3;
345 f[i_coord_offset+DIM*3+YY] += fiy3;
346 f[i_coord_offset+DIM*3+ZZ] += fiz3;
350 fshift[i_shift_offset+XX] += tx;
351 fshift[i_shift_offset+YY] += ty;
352 fshift[i_shift_offset+ZZ] += tz;
355 /* Update potential energies */
356 kernel_data->energygrp_elec[ggid] += velecsum;
357 kernel_data->energygrp_vdw[ggid] += vvdwsum;
359 /* Increment number of inner iterations */
360 inneriter += j_index_end - j_index_start;
362 /* Outer loop uses 41 flops */
365 /* Increment number of outer iterations */
368 /* Update outer/inner flops */
370 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*157);
373 * Gromacs nonbonded kernel: nb_kernel_ElecRF_VdwBham_GeomW4P1_F_c
374 * Electrostatics interaction: ReactionField
375 * VdW interaction: Buckingham
376 * Geometry: Water4-Particle
377 * Calculate force/pot: Force
380 nb_kernel_ElecRF_VdwBham_GeomW4P1_F_c
381 (t_nblist * gmx_restrict nlist,
382 rvec * gmx_restrict xx,
383 rvec * gmx_restrict ff,
384 t_forcerec * gmx_restrict fr,
385 t_mdatoms * gmx_restrict mdatoms,
386 nb_kernel_data_t * gmx_restrict kernel_data,
387 t_nrnb * gmx_restrict nrnb)
389 int i_shift_offset,i_coord_offset,j_coord_offset;
390 int j_index_start,j_index_end;
391 int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
392 real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
393 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
394 real *shiftvec,*fshift,*x,*f;
396 real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
398 real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
400 real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
402 real ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
404 real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
405 real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
406 real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
407 real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
408 real dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
409 real velec,felec,velecsum,facel,crf,krf,krf2;
412 real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
421 jindex = nlist->jindex;
423 shiftidx = nlist->shift;
425 shiftvec = fr->shift_vec[0];
426 fshift = fr->fshift[0];
428 charge = mdatoms->chargeA;
432 nvdwtype = fr->ntype;
434 vdwtype = mdatoms->typeA;
436 /* Setup water-specific parameters */
437 inr = nlist->iinr[0];
438 iq1 = facel*charge[inr+1];
439 iq2 = facel*charge[inr+2];
440 iq3 = facel*charge[inr+3];
441 vdwioffset0 = 3*nvdwtype*vdwtype[inr+0];
446 /* Start outer loop over neighborlists */
447 for(iidx=0; iidx<nri; iidx++)
449 /* Load shift vector for this list */
450 i_shift_offset = DIM*shiftidx[iidx];
451 shX = shiftvec[i_shift_offset+XX];
452 shY = shiftvec[i_shift_offset+YY];
453 shZ = shiftvec[i_shift_offset+ZZ];
455 /* Load limits for loop over neighbors */
456 j_index_start = jindex[iidx];
457 j_index_end = jindex[iidx+1];
459 /* Get outer coordinate index */
461 i_coord_offset = DIM*inr;
463 /* Load i particle coords and add shift vector */
464 ix0 = shX + x[i_coord_offset+DIM*0+XX];
465 iy0 = shY + x[i_coord_offset+DIM*0+YY];
466 iz0 = shZ + x[i_coord_offset+DIM*0+ZZ];
467 ix1 = shX + x[i_coord_offset+DIM*1+XX];
468 iy1 = shY + x[i_coord_offset+DIM*1+YY];
469 iz1 = shZ + x[i_coord_offset+DIM*1+ZZ];
470 ix2 = shX + x[i_coord_offset+DIM*2+XX];
471 iy2 = shY + x[i_coord_offset+DIM*2+YY];
472 iz2 = shZ + x[i_coord_offset+DIM*2+ZZ];
473 ix3 = shX + x[i_coord_offset+DIM*3+XX];
474 iy3 = shY + x[i_coord_offset+DIM*3+YY];
475 iz3 = shZ + x[i_coord_offset+DIM*3+ZZ];
490 /* Start inner kernel loop */
491 for(jidx=j_index_start; jidx<j_index_end; jidx++)
493 /* Get j neighbor index, and coordinate index */
495 j_coord_offset = DIM*jnr;
497 /* load j atom coordinates */
498 jx0 = x[j_coord_offset+DIM*0+XX];
499 jy0 = x[j_coord_offset+DIM*0+YY];
500 jz0 = x[j_coord_offset+DIM*0+ZZ];
502 /* Calculate displacement vector */
516 /* Calculate squared distance and things based on it */
517 rsq00 = dx00*dx00+dy00*dy00+dz00*dz00;
518 rsq10 = dx10*dx10+dy10*dy10+dz10*dz10;
519 rsq20 = dx20*dx20+dy20*dy20+dz20*dz20;
520 rsq30 = dx30*dx30+dy30*dy30+dz30*dz30;
522 rinv00 = gmx_invsqrt(rsq00);
523 rinv10 = gmx_invsqrt(rsq10);
524 rinv20 = gmx_invsqrt(rsq20);
525 rinv30 = gmx_invsqrt(rsq30);
527 rinvsq00 = rinv00*rinv00;
528 rinvsq10 = rinv10*rinv10;
529 rinvsq20 = rinv20*rinv20;
530 rinvsq30 = rinv30*rinv30;
532 /* Load parameters for j particles */
534 vdwjidx0 = 3*vdwtype[jnr+0];
536 /**************************
537 * CALCULATE INTERACTIONS *
538 **************************/
542 c6_00 = vdwparam[vdwioffset0+vdwjidx0];
543 cexp1_00 = vdwparam[vdwioffset0+vdwjidx0+1];
544 cexp2_00 = vdwparam[vdwioffset0+vdwjidx0+2];
546 /* BUCKINGHAM DISPERSION/REPULSION */
547 rinvsix = rinvsq00*rinvsq00*rinvsq00;
548 vvdw6 = c6_00*rinvsix;
550 vvdwexp = cexp1_00*exp(-br);
551 fvdw = (br*vvdwexp-vvdw6)*rinvsq00;
555 /* Calculate temporary vectorial force */
560 /* Update vectorial force */
564 f[j_coord_offset+DIM*0+XX] -= tx;
565 f[j_coord_offset+DIM*0+YY] -= ty;
566 f[j_coord_offset+DIM*0+ZZ] -= tz;
568 /**************************
569 * CALCULATE INTERACTIONS *
570 **************************/
574 /* REACTION-FIELD ELECTROSTATICS */
575 felec = qq10*(rinv10*rinvsq10-krf2);
579 /* Calculate temporary vectorial force */
584 /* Update vectorial force */
588 f[j_coord_offset+DIM*0+XX] -= tx;
589 f[j_coord_offset+DIM*0+YY] -= ty;
590 f[j_coord_offset+DIM*0+ZZ] -= tz;
592 /**************************
593 * CALCULATE INTERACTIONS *
594 **************************/
598 /* REACTION-FIELD ELECTROSTATICS */
599 felec = qq20*(rinv20*rinvsq20-krf2);
603 /* Calculate temporary vectorial force */
608 /* Update vectorial force */
612 f[j_coord_offset+DIM*0+XX] -= tx;
613 f[j_coord_offset+DIM*0+YY] -= ty;
614 f[j_coord_offset+DIM*0+ZZ] -= tz;
616 /**************************
617 * CALCULATE INTERACTIONS *
618 **************************/
622 /* REACTION-FIELD ELECTROSTATICS */
623 felec = qq30*(rinv30*rinvsq30-krf2);
627 /* Calculate temporary vectorial force */
632 /* Update vectorial force */
636 f[j_coord_offset+DIM*0+XX] -= tx;
637 f[j_coord_offset+DIM*0+YY] -= ty;
638 f[j_coord_offset+DIM*0+ZZ] -= tz;
640 /* Inner loop uses 139 flops */
642 /* End of innermost loop */
645 f[i_coord_offset+DIM*0+XX] += fix0;
646 f[i_coord_offset+DIM*0+YY] += fiy0;
647 f[i_coord_offset+DIM*0+ZZ] += fiz0;
651 f[i_coord_offset+DIM*1+XX] += fix1;
652 f[i_coord_offset+DIM*1+YY] += fiy1;
653 f[i_coord_offset+DIM*1+ZZ] += fiz1;
657 f[i_coord_offset+DIM*2+XX] += fix2;
658 f[i_coord_offset+DIM*2+YY] += fiy2;
659 f[i_coord_offset+DIM*2+ZZ] += fiz2;
663 f[i_coord_offset+DIM*3+XX] += fix3;
664 f[i_coord_offset+DIM*3+YY] += fiy3;
665 f[i_coord_offset+DIM*3+ZZ] += fiz3;
669 fshift[i_shift_offset+XX] += tx;
670 fshift[i_shift_offset+YY] += ty;
671 fshift[i_shift_offset+ZZ] += tz;
673 /* Increment number of inner iterations */
674 inneriter += j_index_end - j_index_start;
676 /* Outer loop uses 39 flops */
679 /* Increment number of outer iterations */
682 /* Update outer/inner flops */
684 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*39 + inneriter*139);